ln-812-optimization-researcher

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). The skill's mandatory workflow (Phase 1 and Research Protocol) explicitly directs the agent to use WebSearch and WebFetch (including fetching missing shared files from raw.githubusercontent.com) to ingest public web content for industry/competitor research, meaning untrusted third-party content is read and used to drive decisions and hypotheses.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly states that if shared/ is missing it will fetch required instruction files at runtime from https://raw.githubusercontent.com/levnikolaevich/claude-code-skills/master/skills/{path}, and those fetched files (e.g., tool-priority and mandatory-read docs) directly control the agent's prompt/behavior, so this is a high-risk runtime dependency.